Systems and methods for identifying correlations of certain scenarios to performance of network communications

Abstract

Systems and methods are provided for receiving a set of feature vectors. Each feature vector in the set may comprise feature values for a plurality of features associated with network communications. A first score for a first subset of the feature vectors that have at least one common feature value for a first feature of the plurality of features may be determined. A second score for a second subset of the feature vectors may be determined. The second subset may comprise the first subset and other feature vectors that have a different feature value for the first feature. Based on a change between the first score and the second score, whether to group the common feature value and the different feature value together may be determined.

What is claimed is: 
     
       1. A method comprising:
 receiving, by a correlation determination engine, a set of feature vectors associated with a performance metric, wherein each feature vector in the set comprises feature values for a plurality of features associated with network communications; 
 determining, by the correlation determination engine, a first score for a first subset of the feature vectors that have at least one common feature value for a first feature of the plurality of features; 
 determining, by the correlation determination engine, a second score for a second subset of the feature vectors, the second subset comprising the first subset and other feature vectors that have a different feature value for the first feature; 
 based on a change between the first score and the second score, determining, by the correlation determination engine, whether to group the common feature value and the different feature value together, 
 wherein the common feature value and the different feature value help identify a dominant scenario associated with success or failure conditions of the network communications; and 
 providing, by the correlation determination engine, a group of feature values comprising the common feature value and the different feature value as a set of feature values that has greater correlation with the performance metric than other groups of feature values that exclude the different feature value, upon determining to group the common feature value and the different feature value. 
 
     
     
       2. The method of  claim 1 , further comprising:
 determining that the second score is greater than the first score; and 
 grouping the common feature value and the different feature value together. 
 
     
     
       3. The method of  claim 2 , further comprising:
 generating a recommendation of a remedial action based on the group of feature values. 
 
     
     
       4. The method of  claim 1 , wherein the determining the first score for the first subset of the feature vectors comprises:
 calculating the first score based on a failure rate and a failure fraction associated with the common feature value, wherein the failure rate is defined based on a total number of failures in the first subset in relation to a total number of attempts in the first subset and the failure fraction is defined as a total number of failures in the first subset in relation to the total number of failures for the network communications. 
 
     
     
       5. The method of  claim 4 , wherein the first score is a harmonic mean of the failure rate and the failure fraction. 
     
     
       6. The method of  claim 5 , wherein the harmonic mean is adjusted with a weight applied to at least one of the failure rate or the failure fraction. 
     
     
       7. The method of  claim 1 , further comprising:
 for each feature value associated with the first feature, determining a corresponding subset of feature vectors that have the feature value for the first feature; and 
 determining a corresponding score for the feature value. 
 
     
     
       8. The method of  claim 7 , further comprising:
 sorting the feature value in a sorted list of feature values associated with the first feature based on the corresponding score. 
 
     
     
       9. The method of  claim 8 , wherein the common feature value is selected based on a position of the common feature value in the sorted list and the different feature value is selected based on a different feature value having the next position to the position in the sorted list. 
     
     
       10. The method of  claim 1 , further comprising:
 determining a filter comprising at least the common feature value that, when applied to the set of feature vectors, results in the first subset. 
 
     
     
       11. The method of  claim 10 , further comprising:
 upon determining that the second score is greater than the first score, updating the filter to include the different feature value. 
 
     
     
       12. The method of  claim 11 , wherein the feature vectors are associated with a performance metric, the method further comprising:
 providing a group of feature values comprising the common feature value and the different feature value as a set of feature values that has greater correlation with the performance metric than other groups of feature values that do not comprise the common feature value and the different feature value; 
 apply the filter to the set of the feature vectors to determine a filtered set empty of feature vectors associated with the filter; and 
 providing a second group of feature values remaining in the filtered set as a second set of feature values that has greater correlation with the performance metric than other groups of feature values in the filtered set that do not comprise feature values in the second group of feature values. 
 
     
     
       13. The method of  claim 12 , further comprising:
 determining a third score for a third subset of the feature vectors that have at least one second common feature value for a second feature of the plurality of features; and 
 based on a change between the second score and the third score, determining whether to update the filter to include the second common feature value. 
 
     
     
       14. The method of  claim 1 , wherein the group of feature values identifies a scenario, the method further comprising:
 querying a data store for additional data associated with at least one feature vector defined by at least one feature value of the group of feature values; and 
 providing the additional data in association with the group of feature values. 
 
     
     
       15. The method of  claim 1 , wherein the performance metric comprises at least one of a binary feature value and a non-binary feature value. 
     
     
       16. The method of  claim 1 , wherein a feature of a feature vector is associated with a continuous range of feature values, the method further comprising:
 translating feature values of the feature to a discrete range of feature values based on statistics associated with peer networks for the feature. 
 
     
     
       17. The method of  claim 1 , wherein the determining the first score for the first subset of the feature vectors comprises:
 calculating the first score based on a success rate and a success fraction associated with the common feature value, wherein the success rate is defined based on a total number of successes in the first subset in relation to a total number of attempts in the first subset and the success fraction is defined as a total number of successes in the first subset in relation to the total number of successes for the network communications. 
 
     
     
       18. A system comprising:
 a processor; and 
 a non-transitory storage medium storing instructions that, when executed on the processor, performs a method comprising: 
 receiving a set of feature vectors associated with a performance metric, wherein each feature vector in the set comprises feature values for a plurality of features associated with network communications; 
 determining a first score for a first subset of the feature vectors that have at least one common feature value for a first feature of the plurality of features; 
 determining a second score for a second subset of the feature vectors, the second subset comprising the first subset and other feature vectors that have a different feature value for the first feature; and 
 based on a change between the first score and the second score, determining whether to group the common feature value and the different feature value together, 
 wherein the common feature value and the different feature value help identify a dominant scenario associated successes or failures of the network communications; and 
 providing a group of feature values comprising the common feature value and the different feature value as a set of feature values that has greater correlation with the performance metric than other groups of feature values that exclude the different feature value, upon determining to group the common feature value and the different feature value. 
 
     
     
       19. A non-transitory machine-readable storage medium storing instructions that upon execution cause a system to perform a method comprising:
 receiving a set of feature vectors associated with a performance metric, wherein each feature vector in the set comprises feature values for a plurality of features associated with network communications; 
 determining a first score for a first subset of the feature vectors that have at least one common feature value for a first feature of the plurality of features; 
 determining a second score for a second subset of the feature vectors, the second subset comprising the first subset and other feature vectors that have a different feature value for the first feature; and 
 based on a change between the first score and the second score, determining whether to group the common feature value and the different feature value together, 
 wherein the common feature value and the different feature value help identify a dominant scenario associated successes or failures of the network communications; and 
 providing a group of feature values comprising the common feature value and the different feature value as a set of feature values that has greater correlation with the performance metric than other groups of feature values that exclude the different feature value, upon determining to group the common feature value and the different feature value.